Talk:Polish phonology/Archive 2

Archive 1

Archive 2

Confusing section about ę

Totally confusing, if you permit! What do we learn? gęba is pronounced [ˈɡɛmba]. Nothing "nasal" in there. But hold on...węże is pronounced with a textbook /ɛ̃/ like in French chagrin. Why don't the Polish do a ['wɛnʐe] here too? See, that's where the confusion begins! Lastly: there is absolutely no word of the colloquial language which pronounces się as simply ['ɕʲɛ] - nothing "nasal" either!! As currently, it confuses the learner a lot more than it would benefit him. -andy 217.50.43.11 (talk) 00:36, 24 December 2010 (UTC)

Native speaker here: everything you said is wrong. Gęba is pronounced [ˈgɛ̃mba], with an allophonically nasalized /ɛ/. /ɛ̃, ɔ̃/ before /p, b/ are always pronounced /ɛm, ɔm/. Węże is pronounced [ˈwɛõ̯ʐɛ], with a diphthong with a nasal offglide. This is unlike French, where it is a pure monophthong [ɛ̃] ~ [æ̃] (but in Quebec it can be a nasal diphthong [ẽɪ̃]). This is because before fricatives, /ɛ̃, ɔ̃/ are diphthongs with a nasal second element: [ɛõ̯, ɔõ̯]. Lastly, the last sentence is also not true. [ɕɛ] is the normal pronunciation of się; [ɕɛõ̯] is not wrong, but for me sounds too formal, almost hypercorrect (although it is not according to the standard language). My advice? Never mistake orthography with phonetics. Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 13:30, 8 November 2014 (UTC)

Gęba is pronounced [ˈgɛ̃mba], with an allophonically nasalized /ɛ/.< The allophonic nasalization of vowels before [N] in Polish is a Wikipedia-legend. I've transcribed/analyzed and I'm transcribing/analyzing both colloquial and very controlled Polish recordings. An audible nasalization before [N] is possible but quite rare. And if it happens to a vowel to be nasalized in this position, only a part (max. the half, normally a smaller part) of the vowel is nasalized. And if this partial nasalization is to strong and/or a longer part of the vowel is nasalized, it sounds weird to my ear (I'm a native speaker). -- 2A02:238:F014:6DE:E5E4:58B9:428B:C206 (talk) 17:13, 23 January 2015 (UTC)

I've never said anything about the audibility and duration of the pre-/N/ nasalization of vowels. Let's look at the sources, which vary in their conclusions:

- Gussmann (2007:271) agrees with you; he says that "While nasal diphthongs appear before continuants and word-finally, before a stop we find a sequence of an oral vowel and a nasal homorganic with the stop. Again the vowel shows no traces of nasalization. Consider examples of the basic places of articulation."

- Both Ostaszewska-Trambor (2000) and Rocławski (1976) mention the pre-/N/ nasalization, and Wiśniewski (2001:84) says that it is "possible". The first two were published before the existence of Wikipedia, so your argument of the pre-/N/ nasalization of vowels being a "Wikipedia legend" is completely invalid.

And while your analyses may be interesting, unfortunately they can't serve as sources on Wikipedia - please read WP:OR. Remember to sign your messages. Thanks. — Peter238 (v̥ɪˑzɪʔ mɑˑɪ̯ tˢʰoˑk̚ pʰɛˑɪ̯d̥ʒ̊) 18:40, 23 January 2015 (UTC)

:: Concerning Polish nasal vowels: there's a Polish scholarly publication whose title escapes me at the mo' [shall try to provide] and which says that the tendency to pronounce nasal vowels as a diphthong whose first element is a non-nasal vowel and the second -- an [õ̯], thus 'się' would be [ɕɛõ̯], unless it's [ɕɛ] (which it is more often than not), is a recent, in terms of decades, development. As a senior Polish person I would agree, [ɕɛõ̯] and its ilk strike me as recent, sort of affected, 'genteel', a bit effeminate even. My own pronunciation would be: [ɕɛɛ̯̃] or [wɔɔ̯̃s] (moustache) 2003:E6:3DA:D991:DD74:D094:403D:F3FD (talk) 15:03, 5 November 2018 (UTC) Wojciech Żełaniec

Palatalized velars

I'm at a complete loss for why palatalized velars are listed as separate phonemes, but not other consonants. You can either analyze Polish as having palatalized consonants and then you list them all (so /pʲ/ as in pies /pʲɛs/, /kʲ/ as in kiedy /kʲɛdɨ/, /vʲ/ as in wiać, etc.), or you analyze them as sequences of consonant + /j/ (so you don't list palatalized consonants and pies is transcribed /pjɛs/, kiedy - /kjɛdɨ/, wiać - /vjatɕ/) - as actually mentioned in the article. Then why on Earth does the article list kʲ and gʲ but not pʲ and bʲ? Are you telling me there's a difference in pronunciation of consonant + «i» in pies, wiać and kiedy? Please someone tell me what is going on... — Preceding unsigned comment added by 188.146.224.55 (talk) 00:00, 6 September 2018 (UTC)

Absolutely agree! [kj gj] occur only before [a o u y], and [k g] occur only before [e] [i], so they're in complementary distribution, which means they're allophones, not phonemes. How have they survived in the table for so long? I think the error is so crass and serious that I'm removing them forthwith. Farnwell (talk) 16:25, 27 September 2018 (UTC)

Shouldn't that be the opposite? Do you have a source? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:01, 27 September 2018 (UTC)

I wrote without proper thought (and yes, they're the wrong way round), and you were right to revert the edit. However, I still think that they're pretty inescapably allophones, not phonemes, for three reasons: (1) there's no single-letter spelling that shows the contrast (you have to use two letters); (2) native speakers can't hear the difference; (3) the only breach to complementary distribution is half-a-dozen loanwords beginning with <ke>. I'll post more fully in the next few days. My analysis might be Original Research, and therefore inadmissible on Wikipedia, but I've posted a "low functional load" caveat for this contrast in the article - that at least shouldn't be controversial. - Farnwell (talk) 11:17, 28 September 2018 (UTC)

It still needs a source. I've put a citation tag on the claim in question and removed your signature from the article (which I'm assuming was an accident). I'll give you a few days to provide a source before I remove it. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:05, 28 September 2018 (UTC)

But it doesn't even matter if they have a low functional load or not. /pʲ bʲ/ are not listed as phonemes, so neither should /kʲ gʲ/. Or they should both be listed. It's not a matter of sources, but of logic. Especially given the fact that, as Farnwell mentioned, k-kʲ and g-gʲ are almost in complementary distribution, while p-pʲ and b-bʲ are not (lots of minimal pairs in the case of the latter: bada - biada, buro - biuro, bierze - beże, etc.; no minimal pairs for the former). As for the sources, Polish ones usually list all palatalized consonants as phonemes. For example, Słownik Ortograficzny PWN says: "There are 43 consonants in the modern Polish language: p pʲ b bʲ f fʲ v vʲ m mʲ t tʲ d dʲ n ɲ w l lʲ r rʲ j s ɕ sʲ z ʑ zʲ t͡s t͡ɕ t͡sʲ d͡z d͡ʑ d͡ʒ t͡ʃ ʃ ʒ k kʲ g gʲ x xʲ" (transcription to IPA mine, they use the Slavic notation). For me personally such analysis is unnecessary, but it doesn't matter. It doesn't matter if you transcribe it /pʲes/, /pjes/, /pʲjes/, /pi̯es/ or whatever. The point is, if you transcribe "pies" as /pʲes/, you should transcribe "kiedy" as /kʲedɨ/; and if you transcribe "pies" as /pjes/, you should transcribe "kiedy" as /kjedɨ/. The article as for now mixes analyses. — Preceding unsigned comment added by 188.146.96.83 (talk) 10:44, 29 September 2018 (UTC)

Unfortunately, it must be a matter of sources. You can't apply logic to phonemic inventories like that. The article briefly discusses the rationale for omitting palatalized labials, but there is merit to providing sourcing for this justification, as well as bringing in sources that analyze the language as Słownik Ortograficzny does (kind of like we do for the vowel analyses of Russian at Russian phonology. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:17, 29 September 2018 (UTC)

Nobody transcribes "pies" as /pʲjes/. Whether the sequence [pʲj] (because that's what it is phonetically, the soft character of the first consonant is undisputable) is analyzed as /pʲ/ or /pj/ is an either-or issue. The transcription /pi̯es/ is also pretty much impossible to find in the literature (in sources that use IPA). The transcription of the mid front vowel does vary between ⟨e⟩ and ⟨ɛ⟩ by the way. Kbb2 (ex. Mr KEBAB) (talk) 19:12, 31 August 2019 (UTC)

[w] in impossible (?) positions

What happens to [w] in words like przemysł and łza? Is it even possible to have [w] in this position? Well, for me it isn't, but per se I wonder. The recordings at wiktionary sound to me like [ˈpʃɛmɨsu] and [wuˈza] or perhaps [wɨˈza]. Thank you. — Preceding unsigned comment added by 88.70.71.22 (talk) 18:14, 5 December 2018 (UTC)

This sounds more like an issue of orthography, rather than phonology. The recording for przemysł definitely sounds like [ˈpʂɛmɨsu], but łza sounds like [wza], which might go against sonority hierarchy/Sonority Sequencing Principle, but that's nothing new with Slavic languages. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:48, 5 December 2018 (UTC)

It's probable that the two words don't have exactly the same [w] sound. In przemysł, it's probably slightly longer and vowel-like (EDIT: the way I see it, it's too constricted to be vowel-like, so it's not like Polish [u]), but it's still not an unstressed [u] as in domu. I'd say it's a sound somewhere in-between the typical Polish [w] and [u].

I agree with Aeusoes1 that łza has a canonical [w] sound. Kbb2 (ex. Mr KEBAB) (talk) 20:32, 14 December 2018 (UTC)

I think [ˈpʂɛmɨsw̩] and [ˈpʂɛ.mɨs.w̩] are acceptable narrow transcriptions of przemysł. I think the last consonant should be considered to be syllabic, at least in this pronunciation variant. Kbb2 (ex. Mr KEBAB) (talk) 23:07, 14 December 2018 (UTC)

Thanks a lot! But what is [w̩] anyway? I'd have thought [w̩] to be just another way of writing [u]. What's the difference? And with the other word I didn't mean that it sounded like [uˈza], which it clearly doesn't. But it does sound dissyllabic to me, is it not? — Preceding unsigned comment added by 88.76.16.83 (talk) 00:05, 17 December 2018 (UTC)

By writing it [w̩] I'm trying to convey the fact that it's more close, rounded and tense than Polish [u] which clearly isn't the same as the cardinal [u] (see cardinal vowels). Polish [u] is something in-between the canonical values of the IPA signs ⟨u⟩ and ⟨ʊ⟩ (the latter isn't a cardinal vowel though, it's something in-between cardinal [u], [ʉ] and [o]), though it is closer to cardinal [u]. This is one of the reasons we have so much trouble differentiating /uː/ and /ʊ/ in Standard German - we produce the former as too short, insufficiently rounded, somewhat too open and too lax, whereas our attempts at SG /ʊ/ end up to be too close and perhaps too strongly rounded (German /ʊ/ has consistent weak rounding, whereas Polish /u/ has a rather variable one).

I would argue that what I write [w̩] is even closer than Standard German /uː/, which is pretty much a perfect cardinal [u]. I'd say it has the phonetic quality of an approximant and the length of an unstressed vowel. So even in fully narrow IPA [w̩] and [u] would probably not denote exactly the same sounds.

I know. I think neither of us (at least I didn't) implied that you consider that recording of łza to sound like [uˈza]. I speak Polish natively (not that it's very important here, just saying), and to me the word doesn't sound disyllabic. As I've already said, to me [w] in that word sounds exactly the same as the canonical labiovelar approximant in words such as ładny. I hear no difference between them. Kbb2 (ex. Mr KEBAB) (talk) 01:12, 17 December 2018 (UTC)

It acts like one syllable, even if it sounds like there may be a hint of a second. (Not a native speaker, but know some of the language.) In singing, after a vowel it might sometimes sound like an elision with the previous vowel and /u/, but it will still be one syllable. You can hear this in Aleksandra Kurzak's recording of Chopin's Polish songs (she takes some of them, Mariusz Kwiecień the others): in Op. 74 No. 2 (Wiosna) we have two consecutive lines "W oku łza się kręci. / Łza wybiegła z oka!", and you can hear this happen between the lines. A similar thing happens with things like the n in piosnka (possible poetic form of piosenka "song", appears in the same poem); it is not treated as a separate syllable, though it may sound like a short one is briefly present. I think this is similar to the phenomenon John Wells described on his blog about names like Gdynia and Gdańsk:

“ It wasn’t until I first visited Poland, when I was twenty, that I discovered that in Polish the initial plosives in Gdańsk ɡdaj̃sk and Gdynia ˈɡdɨɲa are not like the English gd in hugged. Rather, when my Polish friends demonstrated the pronunciation to me, they released the velar plosive BEFORE completing the approach for the dental. This inevitably gave rise to a tiny transitional vocoid between the two hold phases, but it was not long enough to count as a separate schwa segment (compare the definite schwa in English Godiva). ”

Double sharp (talk) 14:29, 5 January 2020 (UTC)

I've totally forgotten about this thread, but anyway: When I visited Poland in 2019, I actually asked four native speakers to pronounce this word for me. It doesn't really take a trained phonetician to hear [ˈpʂɛmɨsw̩] for przemysł, with three syllables. The colloquial form [ˈpʂɛmɨs] has two syllables, and apparently so does the underlying phonemic structure: /ˈpʂɛmɨsw/. I guess Polish phonology doesn't account for syllabic consonants on a phonetic level (I was also told that tylko is sometimes pronounced [ˈtl̩kɔ], which is very Czech/Slovak to that person's ear).

I agree that this [w̩] is not the same as Polish /u/. Sol505000 (talk) 09:59, 7 March 2021 (UTC)

ŋ

I think the only acceptable situation where [ŋ] appears is as a substitution for the disappearing nasals in Polish like in "dziękuję". However, like in Russian and unlike many other languages, normal N in front of velars like in "bank" shouldn't turn into a velar [ŋ]. --81.18.214.118 (talk) 22:47, 25 August 2019 (UTC)

• Eee, I and everyone I know pronounce "bank" and every single word ending in "-nk/ng" with a velar [ŋ]. Stop spreading misinformation. Shumkichi (talk) 03:42, 13 July 2021 (UTC)

Affricates and stop–fricative clusters

Moved from User talk:Nardog

I can see that you reverted the edit made by Mazab IZW. I agree that these transcriptions [ˈt͡ʂtʂɨsta], [ˈd͡ʐdʐɛm] are wrong. However, Zagórska points out earlier in her paper that the cluster [tʂ] has two or three elements: [t]₂, an optional release of [t]₂ noted by [ʰ] and [ʂ]₂. This facet and her conclusion that [ʂ] is longer in clusters give either [tʰʂː] or [tʂː]. Other descriptions (e.g. Sawicka (1995:150), Ostaszewska & Tabor (2000:87)) dub [tʰʂː] as [t̺ʂ] and interpret no release as assimilation of [t] to [t͡ʂ], so [tʂː] is transcribed as [t͡ʂʂ]. My proposal:

Polish contrasts affricates and stop–fricative clusters by the fricatives being consistently longer in clusters than in affricates:

• [ˈt͡ʂɨsta] ("clean" fem.) vs. [ˈtʰʂːɨsta] or [ˈtʂːɨsta] ("three hundred").
• [ˈd͡ʐɛm] ("jam") with vs. [ˈdʱʐːɛm] or [ˈdʐːɛm] or ("take a nap" imper. sing.).

A similar convention to differentiate between affricates from clusters by using [t͡s, d͡z] vs [tʰs, dʱz] is used by is used by Rybka (2015:89). Please let me know your thoughts.

--Qerez (talk) 16:53, 19 September 2021 (UTC)

While Zagórska Brooks does report aspiration in some utterances, her conclusion is that "the relative length of [š] was found to be the only consistent element of distinction between [tš] and [t͜š]", so it makes little sense to me to show the transcription with aspiration first. Also the full length mark ⟨ː⟩ usually connotes the segment is twice as long, while Zagórska Brooks reports fricatives in clusters are almost exactly 1.5 times longer than those in affricates, so the half-length mark ⟨ˑ⟩ makes a lot more sense. Finally, the transcriptions in the bullet points aren't meant to convey "conventions" but rather to illustrate the actual phonetic differences in IPA form. It indeed doesn't hurt to mention that such a convention exists (as much as I find it misleading because it suggests aspiration is consistently present in clusters, which it isn't according to Zagórska Brooks), but I'd scratch the word "similar". Nardog (talk) 06:15, 20 September 2021 (UTC)

I'd prefer the version with release to be listed first. This is considered by Sawicka and others as the unassimilated version, a realization of /tʂ/). The version with no audible release is considered a realization of an assimilation, affrication of the segment /t/: /tʂ/ → /t͡ʂʂ/. I'll change the length mark. I think I will write a bit longer paragraph. Thanks for constructive feedback.

--Qerez (talk) 07:41, 20 September 2021 (UTC)

A stop with "no audible release" before a continuant would mean there's a pause between the stop and fricative. I don't think that's what you meant there, so I fixed that. A problem I see with listing the transcriptions with aspiration first is that in Zagórska Brooks's study, only three out of eight tokens had aspiration. Are you sure this is not something only found in careful speech? Do the other sources you cited back up the claim by empirical studies? Nardog (talk) 11:46, 20 September 2021 (UTC)

Both audio examples given in the article feature aspiration so listing them first make more sense from that perspective. "Trzysta" sounds spontaneous, while "drzem" sounds a bit exaggerated. Prescriptively, both are correct both in careful and colloquial speech. Dunaj (2006:170) in Zasady poprawnej wymowy polskiej writes:

"Połączenia TRZ, DRZ występują w nagłosie, śródgłosie i wygłosie (-TRZ) wyrazów. W wymowie odpowiadają im trzy realizacje a) tsz, d-ż, b) czsz, dżż, b) cz, dż. W wymowie staranej, a także potocznej dopuszczone są w nagłosie i śródgłosie dwa typy wymowy: tszoda // czszoda, d-żewo, dżżewo, pojutsze // pojuczsze. Drugi typ dominuje w odmianie potocznej. Realizacje uproszczone (z zanikiem spółgoski szczelinowej), a więc czoda, dżewo, wystepujące często, szczególnie w niektórych regionach Polski, są niepoprawne zarowno w polszczyźnie starannej, jak i potocznej. W wygłosie w formach takich jak patrz, popatrz w wymowie staranej poprawną realizacją jest grupa tsz, por. patsz, popatsz (podawane w niektorych opracowaniach formy paczsz, popaczsz w rzeczywistości nie wystepują). W wymowie potocznej obok realizacji patsz, popatsz dopuszcza się wymowę z grupą uproszczoną, a więc pacz, popacz.

["The combinations of TRZ, DRZ occur word initial, medial and final positions (-TRZ). In the pronunciation, they correspond to three realizations: a) tsz, d-ż, b) czsz, dżż, b) cz, dż. Two types of pronunciation are allowed word initially and medially in careful pronunciation, but equally in colloquial one: tszoda // czszoda, d-żewo, dżżewo, pojutsze // pojuczsze. The second type dominates in colloquial speech. Simplified realizations (without a fricative) czoda, dżewo, often occurring, especially in some regions of Poland, are incorrect both in careful and colloquial Polish. In careful speech, forms such as patrz, popatrz are correctly realized by the group tsz, cf. patsz, popatsz (forms paczsz, popaczsz given in some descriptions do not actually appear). In colloquial pronunciation, in addition to patsz, popatsz, pronunciation with a simplified group is allowed, i.e.pacz, popacz."]

In spontaneous speech production I hear both in free variation. In this recording by Wojtula you can hear both in one phrase. I've added a version produced by me (yazoo).

The assimilation is not limited to /tʂ, dʐ/, it occurs for other TS clusters e.g. o ciebie /ɔt͡ɕɛbjɛ/ "about you" vs od siebie /ɔtɕɛbjɛ/ "from each other", octu /ɔt͡stu/ "vinegar" (gen. sg.) vs od stu /ɔtstu/ "from hundred".

Dukiewicz indicates her own paper "Polskie diady spółgłoskowe typu TS i ST" ("Polish consonantal dyads of type TS and ST") but I cannot get hold of it.

The phenomenon seems to be very similar to affrication of /t, d/ in /tr, dr/ by some speakers of English who may produce [tɹ̝̊ɹ̥ , dɹ̝ɹ] or [tʂɻ̊, dʐɻ].

What about using the "unaspirated" diacritic of the Extended IPA to instead "aspiration" as suggested in No audible release?

--Qerez (talk) 19:28, 20 September 2021 (UTC)

The symbol you're looking for is the superscript ⟨ᵊ⟩ (so that trzysta is pronounced [tᵊˈʂɨsta], or perhaps with a voiceless schwa instead. In drzewo, the schwa is voiced). The "unaspirated" diacritic is inappropriate as all stops in Polish are unaspirated. Sol505000 (talk) 07:38, 21 September 2021 (UTC)

Yes, ⟨ᵊ⟩ would imply it's voiced so it's not a good fit. I don't see anything wrong with ⟨ʰ, ʱ⟩. Aspiration is a matter of degree so it's okay to use them as long as the context calls for them, even if it's shorter than phonemic aspiration in other languages, just like we're using ⟨ˑ⟩ only in this particular context to illustrate the difference. The tie bar already indicates assibilated release, so I don't see a need to modify the affricates. Nardog (talk) 08:39, 21 September 2021 (UTC)

But ⟨ʱ⟩ implies a murmured voice, rather than a simple voicing like ⟨ᵊ⟩ does. I'd rather use ⟨ʰ, ᵊ⟩. Sol505000 (talk) 14:57, 25 April 2022 (UTC)

References

Velar nasal as a separate phoneme

The article says that the velar nasal is an allophone of /n/ in front of velars. Edmund Gussman in the Phonology of Polish (p.12) argues for the velar nasal as a separate phoneme, mentioning that minimal pairs such as łąka [wɔŋka] 'meadow' and łonka [wɔnka] 'bosom, dim. nom. pl.' can be found. He does mention that the velar nasal has a highly restrictive distribution, appearing only before velar plosives and that the second word is morphologically complex, which shouldn't matter as morphology is usually seen as independent of phonology. Thoughts? — Preceding unsigned comment added by 81.241.31.169 (talk) 22:15, 20 June 2018 (UTC)

There are probably ways of analyzing that phoneme away, but it doesn't seem like we can call it an allophone of /n/ if [n] can appear in that context. I vaguely recall a source saying that velar assimilation of nasals is not really a thing in Polish (it isn't in Russian except for a few loanwords). — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:35, 21 June 2018 (UTC)

@Aeusoes1: Eee, what source would claim such stupidity? Nasal //N// before velar k, g, and h gets velarised to [ŋ] ALWAYS. The article the guy above mentions is only about the dying North-Eastern dialects near the border with Lithuania and Belarus where they don't velarise their "n". However, nowadays, one can only hear [ŋ] in such positions in standard Polish. Shumkichi (talk) 03:49, 13 July 2021 (UTC)

That's not true. Piosenka varies between /pjɔˈsɛnka/ and /pjɔˈsɛŋka/ in Polish. I'm using phonemic slashes here since the difference between /n/ and /ŋ/ is phonemic in Polish, see Jassem 2003 and Gussmann 2007. Sol505000 (talk) 16:28, 25 April 2022 (UTC)

Vowel heights

Are we sure it is 4x(2:1:2:1)? 2x3(close&close-mid merged; open-mid&open merged) is a simpler explanation of Polish vowel system. Erkinalp9035 (talk) 16:26, 19 September 2018 (UTC)

If I understand you correctly, I agree. Does my recent edit to the vowel chart reflect your thinking? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:30, 19 September 2018 (UTC)

If /ɨ/ belongs with the close series /i, u/, then /a/ belongs with the open series /ɛ, ɔ/. The differences in height are basically the same, the peripheral vowels are about two heights closer than the central ones. Sol505000 (talk) 16:30, 25 April 2022 (UTC)

Soft consonants not included in article

On consonant table are soft versions of "ch", "g" and "k", but there are missing soft "m", "b", "p", "d", "t", "w", "f" and "l". Maybe I'm wrong, but I think this is mistake. Szmomsz (talk) 08:26, 28 August 2019 (UTC)

The table lists just phonemes. The soft variants of those consonants are conditional allophones, not separate phonemes. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:23, 28 August 2019 (UTC)

So why in article about russian phonology they are included? In IPA they are the same. Szmomsz (talk) 12:38, 31 August 2019 (UTC)

@Szmomsz: Because in Russian they are phonemes. In Polish they aren't, at least not according to most of the recent analyses. Kbb2 (ex. Mr KEBAB) (talk) 13:04, 31 August 2019 (UTC)

Thank you very much Szmomsz (talk) 14:19, 31 August 2019 (UTC)

I mean why soft "h", "g", and "k" are recognized as phonemes Szmomsz (talk) 14:35, 31 August 2019 (UTC)

Usually, the analysis rests on the presence of minimal pairs, or sets of words that are distinguished by only one sound. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:56, 2 September 2019 (UTC)

What minimal pairs would distinguish soft "h", "g", and "k" from their plain counterparts? I am still unconvinced as to why there is a palatalized velar series without an accompanying palatalized labial series. For instance, the Wikipedia article has a table which shows that plain "k" is found in kmin /kmin/ and that palatalized "k" is found in kiedy /kʲɛdɨ/; I can make the same argument that plain "p" is found in pas /pas/ and that palatalized "p" is found in pięć /pʲɛ̃t͡ɕ/. In truth, I would go the other way and suggest that instead of having both a palatalized velar series and a palatalized labial series, that both series (including the palatalized velar series) should not be included as Polish phonemes for a reason in the next sentence. Unlike the consonants in the palatal series (which themselves are the palatalized versions of the alveolar series /t/→/t͡ɕ/, /d/→/d͡ʑ/, /s/→/ɕ/, /z/→/ʑ/) and /l/ (which is pointed out in this talk page to historically be the palatalized version of the current /w/), a palatalized velar or palatalized labial consonant can't precede another consonant nor end a word. Hence you can have words like pas /pas/, paś /paɕ/, ukośnik /ukɔɕɲik/, chciwość /xt͡ɕivɔɕt͡ɕ/, and mam /mam/ but something like */ukɔɕɲikʲ/, */xʲt͡ɕivɔɕt͡ɕ/ or */mamʲ/ are impossible. Because of that, there is no (need to make a) separate letter/digraph for palatalized velars and labials, as they don't contrast with their plain counterparts, nor with their plain counterparts + /j/. If pięć is transcribed as /pjɛ̃t͡ɕ/, it would only be valid to transcribe kiedy as /kjɛdɨ/ for consistency's sake.

TLDR: As far as I know, palatalized velars and palatalized labials share the exact same phonological properties that doesn't make them contrastable to a plain velar/labial + /j/ sequence, despite the fact that only the palatalized velars are phonemic according to Wikipedia. I believe that there is an error that should be corrected on the main page, or that an explanation on the main page should be inserted to explain the phonemicity of the palatalized velars. Thank you. Excusememoi (talk) 04:49, 28 October 2019 (UTC)

Our policy on not doing your own research/analysis says to rely instead on reliable sources. In this case, at least one of the sources cited in the article provides word lists to show contrasts. Jassem (2003), who doesn't consider soft [xʲ] to be a separate phoneme, provides the near-minimal pairs kielnia ('a trowel') vs. kelner ('waiter'), drogie ('dear [nom. pl.]' vs drogę 'way [acc. sing.]'). I don't have access to Rocławski (1976), so I don't know if that source provides relevant minimal pairs for the soft/hard velar fricative.

Looking through the article history, I can see that Jassem (2003) was cited as the source for the table (with /xʲ/ listed as a phoneme, even though this source does not say as much). Then the citation to Jassem was removed and Ahls23 added the citation to Rocławski (1976) but turned the table into one that included allophones. Kbb2 converted the table to its current form without removing xʲ. So perhaps either of these two editors can provide clarification on the matter. If not, we can restore the citation to Jassem (2003) and conform the table to what's in that source. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:57, 28 October 2019 (UTC)

@Aeusoes1: This is how the table should look like. AFAICS the close central vowel is clearly non-phonemic, and its allophonic status is (to me at least) even clearer than it is in Russian (where it can occur word-initially - in Polish, it can't). Thoughts? I based most of it on Gussmann and The Slavonic Languages. Kbb2 (ex. Mr KEBAB) (talk) 13:09, 29 October 2019 (UTC)

If there is scholarly dispute about the phonemicity of ⟨y⟩, then it would make sense to include all of those sounds, though perhaps marking the ones that would be considered allophones if ⟨y⟩ represents an independent phoneme. Putting them in parentheses is an easy option, though we could also have fun and mess with colors. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:45, 29 October 2019 (UTC)

I would love a table plus explanation like what you gave to appear in the main article. I was initially surprised by the amount of soft consonants that you suggested, but you did explain that you're also accounting consonants only found in loanwords. In this case, it would good to have the loanword-only phonemes be complemented with an asterisk/footnote or in parentheses to show that. I was also surprised to see that /l/ is also a hard consonant as well as a soft one, but you have explained how it was historically considered soft but now reinterpreted as hard by modern speakers, which explains why /w/ is a liquid as well as a glide. You even included an explanation saying that palatal consonants are the only soft consonants to exist word-final and that all other soft consonants occur in other positions and can instead be analyzed as allophones of hard consonants following /j/ (preceding a vowel that's not [i] or [ɨ]) or /i/. Excusememoi (talk) 23:52, 29 October 2019 (UTC)

@Kbb2: "Clearly non-phonemic" - honey, pls, do you understand what the word "phoneme" means? [ɨ] is a separate phoneme as it contrasts with [i] but ONLY in open syllables after nasal consonants. In all the other environments, the two are allophones, albeit in free variation so that they may appear phonemic in such cases, hence the confusion. Also, the most recent borrowings from English have enabled the consonants that were always followed only by [i] or only by [ɨ] to be also followed by the other vowel (e.g. czipsy, tir). This is one of the arguments made by authors who claim that [ɨ] is a phoneme in ALL environments (I disagree with them, though). Shumkichi (talk) 22:31, 11 July 2021 (UTC)

First of all, I'd refraing from using such patronizing language. I don't know what is meant by "In all the other environments, the two are allophones, albeit in free variation so that they may appear phonemic in such cases, hence the confusion." What you're saying is at odds with every source that I've read. I don't recall any source that considers [ɨ] to be non-phonemic (maybe The Slavonic Languages does that, I'm not sure). Why on earth would my be /mɨ/ but ty /ti/? You said it yourself that tir is pronounced [tir], so this makes zero sense! I think it is you who needs to learn the definition of a phoneme. Sol505000 (talk) 16:37, 25 April 2022 (UTC)

Pronunciation of the alveolo-palatals and y

While the article argues that ć, ś, ź, and dź are pronounced /t͡ɕ/, /ɕ/, /ʑ/, and /d͡ʑ/, respectively. However, any native Polish speaker will tell you that this is not how these letters are realised, and it is pretty evident that their pronunciation is distinct from that of Russian ч, щ, жь (as in позже), and джь, respectively. Instead, the Polish set is softer than the Russian set is pronounced more like /t͡ɕsʲ/, /ɕsʲ/, /ʑzʲ/, and /d͡ʑzʲ/ than what the article argues they are pronounced.

I have an additional issue with the transcription of y as /ɨ/, which is only the correct transcription when y is in between two consonants, while in any other case it's pronounced closer to /ɪ/ or even /ʏ/, but this issue was already raised before, so I won't bother to raise it again. YourAverageMax (talk) 23:28, 11 August 2020 (UTC)

Do you have sources to back up your claims? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 01:40, 12 August 2020 (UTC)

If I had any sources, I'd go ahead and implement the changes myself. However, if we look at Wikimedia recordings of Polish words by native Poles, we can already see that the current transcription is inaccurate: for example, compare and ), where the last consonant is different between the two words, although even that's not the best example. I'm sure there should be sources which describe this phenomenon, but I just can't find them. YourAverageMax (talk) 11:37, 12 August 2020 (UTC)

Then also take a listen to any of the audio files provided in this link, and tell me whether Polish ć is closer to Slovak /c/ or Russian /t͡ɕ/. To me, it seems like many Polish speakers pronounce ć identically to how Slovaks pronounce their soft "t"s, while the rest still pronounce it as a /c/ but with a slightly more palatalised realisation. The same is true for /ɟ/. I don't know what the equivalents of these two would be for ś and ź. YourAverageMax (talk) 11:56, 26 August 2020 (UTC)

No, the Polish alveolo-palatal set is represented correctly in the article as the tongue is in the position between the soft palate and the alveolar ridge, unlike the Czech and Slovak consonants wich are completely palatal. Can people stop doing this folk "lingustics" and look up X-ray tests, for example? Shumkichi (talk) 22:37, 11 July 2021 (UTC)

I have an additional issue with the transcription of y as /ɨ/, which is only the correct transcription when y is in between two consonants, while in any other case it's pronounced closer to /ɪ/ or even /ʏ/, but this issue was already raised before, so I won't bother to raise it again. You have an excellent ear. Polish /ɨ/ is sometimes weakly rounded to [ɵ]. /ɨ/ is a phonemic transcription, so we can use only one symbol for it. ⟨ɨ⟩ was chosen because [ɨ] is the closest cardinal vowel to it. The closest IPA symbol in terms of the actual pronunciation is [ɘ] - or, sometimes, [ɵ], when you're doing a very narrow transcription. As the sounds are often depicted as higher than close-mid on the various vowel charts, ⟨ᵻ ᵿ⟩ may be even better if you're prepared to use non-IPA symbols. Polish ⟨y⟩ is definitely more central than front. It is in-between your typical English KIT vowel [ɪ] and the properly central American schwi sound [ɪ̈] used in the unstressed syllables of words such as wanted, roses, Lenin and Lennon (I forced myself to learn both, so I think that I know what I'm talking about). To my ears, Polish /ɨ/ varies between the latter (especially in running speech) and a slightly fronter sound, but it lacks the definitive spread lips found in the aforementioned English sounds. In that regard it is, AFAIK, much like the Northern German /ɪ/. I'm not aware of any in-depth research that touches on this, but my shallow impression is that /ɨ/ is spread mostly in emphatic speech (probably more so in stressed syllables than those that are unstressed), whereas in running speech the lips are either neutral or slightly rounded. The closest English vowel to it seems to be the Southern English /ʊ/, but even that is not a perfect match (the back rounded [ʊ] used by older speakers and in Multicultural London English is nowhere near the allophonic range of the Polish /ɨ/, it's only the fronter variants that are close to it).

Regarding the alveolo-palatals, I'm not really sure what you're talking about. There's no glide from [ɕ ʑ] to [sʲ zʲ]. We don't even have the latter sounds, except in loanwords (such as sinus [ˈsʲinus] 'sine') before /i/ - and even then, the palatalization is so weak that it is no match for Russian [sʲ zʲ], which historically largely correspond to Polish [ɕ ʑ]. Sol505000 (talk) 16:41, 25 April 2022 (UTC)

j̃

O dear, so many years and Wikipedia keeps claiming that Gdańsk is [ɡdaɲsk] instead of [ɡdaj̃sk] (even though this very page contains an account by John Wells, a mainstream phonetician, to be sure, with the latter pronunciation given), and that gęś is [gɛ̃ɕ], not [gɛj̃ɕ]. Yes, you can speak like this, but it sounds hypercorrect, conceited or dialectal. It might even soon become an old-fashioned pronunciation, the way most people nowadays inadvertently pick up hypercorrect pronunciations from the TV (e.g. jabłko as ['jabwkɔ] instead of ['japkɔ], myśl as [mɨɕl] instead of [mɨɕ], mógł as [mugw] instead of [muk], kłuć as [kwut͡ɕ] instead of [kut͡ɕ], tułów, tułowia as ['tuwuf tu'wovʲa] instead of ['twuf 'twovʲa], niesamowity as [ˌɲɛsamɔ'vʲitɨ] instead of [ˌɲɛsːamɔ'vʲitɨ] etc.), but so far it's still the most natural pronunciation (maybe except after [i], e.g. Mińsk, where the diphthong sounds perhaps too close too [ij], as there used to be Polish dialects with [j] instead of [j̃] or [ɲ] in the syllable coda in at least similar contexts). I'd need too dig up some sources (I've seen at least two while in the varsity), for the time being I'm just leaving a heads-up. 89.64.13.33 (talk) 21:13, 16 October 2020 (UTC)

Hmm, Gdańsk transcribes it with j̃ and this article notes both pronunciations. Is there another place that transcribes in the way that you talk about? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 02:26, 17 October 2020 (UTC)

One source for the gęś as [gɛj̃ɕ] series seems to be https://books.google.pl/books?id=ey8ZAQAAIAAJ&q=szczelinowymi (first hit, strona 55) but I don't have access to it, I can't verify it. 89.64.69.33 (talk) 01:26, 23 September 2022 (UTC)

Eee? Where the hell did you get some of these pronunciations? Yes, "jabłko" is always [japkɔ], but "mógł" is [muk] ONLY in very casual and fast speech; in isolation, I'd always pronounce it [mugw]. "Kłuć" is usually pronounced [kwut͡ɕ], even in fast speech, at least in my region. "Myśl" is also usually [mɨɕl̥]. I've never heard anyone say [twuf] or [ˌɲɛsːamɔ'vʲitɨ] (now, you must've come up with the latter; I'd only say it like that if I wanted to put stress on the word but it would still sound super weird). Stop making things up. Shumkichi (talk) 22:42, 11 July 2021 (UTC)

You were only born in this century, weren't you? Except for "myś" (though I'm not sure at this point), all examples were still standard (spelt standardowy, but still pronounced "standartowy" then) in Warsaw in the nineties (at least among ordinary people, not some high-falluting intelligentsia). I could add "przeczkole" for przedszkole (as sung by children in the intro to Domowe przedszkole).

• Jabłko is no longer "japko" unless you really make an effort to avoid picking up the brave new pronunciation. Same for "gark-" (garnk-)
• "(-)myśl" had to be pronounced without l, or else it's impossible to understand how "(-)krésić" became "(-)kréslić". The contamination must've started while é was still the order of the day & thus both verb groups had identical rimes in the imperative. As to the immediate cause, my bet would be on catechisms and schoolbooks, where imperative sentences abound.
• "Mógł" and other verb forms ending in [-Cw] are mentioned by Perlin (PDF page 14--18).
• According to Jan Tokarski's Traktat o ortografii polskiej (1979), all [(-)wu-] pronunciations are hypercorrect, not just [kwut͡ɕ]. He actually refers to the jocular spelling "Uć" for Łódź as fully phonetic! I'm not sure how accurate his observations were (he was from Podlachia, [w] for ł wasn't his original pronunciation), but if he was right, it's amazing how just a few decades of Prussian-style schooling, mass media exposure and shaming by state-sponsored prescriptivists wreaked havoc on the language (and this is before counting all the dialects that disappeared instead of keeping up getting slowly influenced by the "standard" language).
• As for "tłów", try googling it (along with the quotes). You'll see prescriptivist sites popping up, for a reason.
• "Nessamowytyj" with long s (the word is originally Ukrainian, not Polish) was already noted by Oskar Kolberg in Podolia in the 19th century, no need to make it up (and yes, it was probably caused by frequent emphasis on this word).

Two atrocities more:

• Pojedyńczy is getting ousted by the spelling pronunciation, even though the current spelling ("pojedynczy") resulted from a typo (information from a series of lectures by Zygmunt Saloni).
• Sześćset has been pronounced as if spelt "szejset" for centuries, but the etymological spelling was kept under the influence of sześciuset, sześć etc. Some time ago the prescriptivists thought up a brave new pronunciation: as if spelt "sześset". And has it spread through the media like wildfire! Of course, it's like insisting on mieśski instead of miejski, but miejski isn't spelt mieśćski any longer, allowing for less creativity on the prescriptivists' part. 89.64.69.33 (talk) 00:27, 23 September 2022 (UTC)

Palatal consonants lacking

Why aren't palatal B and P listed as Polish phonemes? They form relevant minimal pairs, for example pasek - piasek, bały - biały. They are not a sequence of non palatalized p and b plus a /j/. 51.175.195.49 (talk) 11:35, 20 December 2021 (UTC)

That depends on analysis. See Polish phonology#Phonological status of palatalized consonants. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:32, 20 December 2021 (UTC)

@Aeusoes1: No, it does not depend on analysis, stop annoying me with your ignorance. Of course they are sequences of a consonant followed by a /j/, lol. Look up X-ray tests, a semivowel can clearly be heard in such cases. Palatalisation is NOT phonemic in Polish, it's only allophonic just as in English. "biały" is not pronounced /ˈbʲa.wɨ/ (try to pronounce the word if you're a native speaker, its duration is longer than in the case of /ˈba.wɨ/ which shouldn't happen if /b/ was JUST palatalised) but [ˈbʲja.wɨ], with the palatalisation of /b/ coming from the following /j/, hence it's just allophonic. Besides, it's a very slight kind of palatalisation, much much much weaker than in Russian but similar to English (as in the word "beak" which in a narrow transcription would be represented as [ˈbʲiːk]). Shumkichi (talk) 21:29, 6 January 2022 (UTC)

If the notion that there might be palatalized labial phonemes in Polish is such an annoyance to you, then you might consider looking into editing the section that I pointed to, since I'm merely sharing what the article says. It would probably be worthwhile exercise to share your knowledge and the relevant sources that back up your claims rather than be rude in the talk page. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:24, 6 January 2022 (UTC)

It does. What he means is phonemic analysis, not narrow phonetic transcription. The pronunciation [ˈbʲjawɨ] can be represented in phonemic transcription as /ˈbjawɨ/ (if you think that the palatalization of /b/ is allophonic) or /ˈbʲawɨ/ (if you think that the palatal glide after /b/ is allophonic) but not /ˈbʲjawɨ/, which is not a valid phonemic transcription (at least not in the analyses I'm aware of). EDIT: Then there's also the analysis /ˈbʲawi/, in which [ɨ] is treated as an allophone of /i/ - but nobody endorses it anymore, AFAIK.

Even Russian белый /ˈbʲelɨj/ features a palatal glide before /e/, it's just that it's shorter/weaker than the /j/ phoneme. This doesn't mean that the analysis /ˈbʲawɨ/ is wrong in Polish - as John Wells puts it, There is not some great phoneme system in the sky from which particular languages select their phonemes, with one IPA symbol always standing for the same thing (source: [1]). Polish /bʲ/ stands for something closer to [bʲj] in comparison with its Russian counterpart (though the latter also features a palatal glide, but not as long/strong as the /j/ phoneme. You could transcribe Russian /bʲ/ as [bʲi̯] or something else and then clarify its phonetic features (that just means length - anything else is pretty much self-evident) in prose, as long as the /j/ phoneme is assigned a separate symbol. This is because Russian contrasts /Cʲ/ with /Cʲj/, believe it or not). The Russian mid back vowel is also closer and considerably more diphthongal/triphthongal in comparison with the Polish /ɔ/, making it sound almost like Polish ło or even łoa, and yet its not uncommon to see both written with ⟨o⟩. I think it's pretty obvious that the word co isn't pronounced [tsʊ̯ɔʌ̯] in Polish, in contrast with Russian что, which is most commonly [ʂtʊ̯ɔʌ̯], though that probably depends to a certain degree on the sentence stress. To add to the confusion, in some transcriptions of Dutch and American English, ⟨o⟩ stands for [oʊ ~ ɵʊ ~ ɔʊ].

English beak is not pronounced [ˈbʲiːk] unless you have a (slight) Slavic accent. English doesn't have allophonic palatalization of stops, except in contact with /j/ and then only in the case of the alveolars (and /m/ [a bilabial nasal stop], according to The Phonetics of English and Dutch). Sol505000 (talk) 09:57, 24 January 2022 (UTC)

Update: on the other hand, writing /Cʲ/ (where the capital letter ⟨c⟩ stands for any consonant apart from the (alveolo-)palatals etc.) does cause some odd differences in phonemic analysis that are counter-intuitive to the natives. Take rozjebać and rozpierdolić (both meaning 'to fuck sth up'), for instance. The former is presumably /rɔˈzjɛbatɕ/ (because jebać 'to fuck' alone is /ˈjɛbatɕ/ and because the only trace of /zʲ/ before /ɛ/ is the consonant /ʑ/, as in ziemia /ˈʑɛmja/ 'soil', historically /ˈzʲɛmʲa/ or something like that), whereas rozpierdolić is /rɔspʲɛrˈdɔlitɕ/ (because pierdolić 'to fuck' alone is /pʲɛrˈdɔlitɕ/). I assume maximal syllable onset (so /rɔˈzjɛbatɕ/, rather than the morphologically transparent /rɔzˈjɛbatɕ/) and that /l/ is by definition 'soft' in this analysis, and any palatalization of it is phonetic (the historical hard l is the contemporary /w/). But phonetically, there is no difference between the alleged palatal approximant in /zj/ in rozjebać and the alleged palatalization of the preceding consonant in /pʲ/ in rozpierdolić - both are actually [Cj], or [Cʲj] if you prefer a narrower transcription: [rɔˈzjɛbatɕ, rɔspjɛrˈdɔlitɕ]. I know that it's not the nicest example but it's perfect for the job - you can easily remember it. I think that, if possible, phonemic transcription should be as accessible to the natives (and advanced L2 speakers) as possible. You shouldn't have to stop and think "hmm... is this /Cj/ or /Cʲ/?" when it literally doesn't matter in Polish. It's not an actual phonemic contrast.

I think that the analysis /Cʲ/ is by analogy with either Russian (the predominant foreign language in Poland in the latter half of the 20th century) or maybe even Lithuanian. In the latter, the orthographic ⟨CiV⟩ actually does stand for a palatalized consonant. Or maybe because Polish used to feature actual palatalized consonants, such as the aforementioned /tʲ, dʲ, sʲ, zʲ/ which correspond to the modern /tɕ, dʑ, ɕ, ʑ/. This is why we have alternations such as mata /ˈmata/ 'a mat' vs. na macie /na ˈmatɕɛ/ 'on a mat'. I think that misspellings such as mjasto for miasto 'city' also contributed to the analysis /Cʲ/ (so /ˈmʲastɔ/), rather than /Cj/ (so /ˈmjastɔ/), as the latter transcription would be basically the same as the misspelling mjasto (which, personally, I have nothing against, it's how the word should be spelled given how it's pronounced - but that's just me), save for the final vowel (often written with a bare ⟨o⟩ anyway - see above). In that case, /ˈmʲastɔ/ would be an attempt to make the IPA transcriptions of Polish look "educated", or even to justify the orthographic distinction between ⟨CjV⟩ and ⟨CiV⟩.

I think that the Masurian variant /ˈmɲastɔ/ mniasto (cf. Czech město /ˈmɲɛstɔ/) also points to the fact that what precedes /a/ is a consonant - especially due to the fact that it can be reduced all the way to /ˈɲastɔ/ niasto!

Ah yes, then there's also the issue of the sixth vowel, or /ɨ/. Those who think that it is an allophone of /i/ will automatically give [tʲ] etc. a phonemic status due to loanwords such as Tico [ˈtʲikɔ] 'Tico' (a car model) and sinus [ˈsʲinus] 'sine', which are pronounced with palatalized dentals, rather than the native alveolo-palatals (in that aspect, Polish is unlike Brazilian Portuguese). But most analyses treat Polish as possessing a six vowel system, in which /ɨ/ occupies the close central position, right in-between the front /i/ and the back /u/ (even though that is at odds with the phonetics of Polish, as natives tend to perceive /ɨ/ as something between /i/ and /ɛ/ (which can be as high and/or as central as /ɨ/, not necessarily resulting in a complete merger), which suggests that Polish has at least three if not four contrastive degrees of openness among the front/central vowels, which probably should be grouped together).

Personally, I see no reason to differentiate between the palatal elements in the syllable onsets of zjeść /zjɛɕtɕ/ 'to eat' (transitive) (or zjeb /zjɛp/ 'fucking retard', if you want something related to the former examples :P) and biały /ˈbjawɨ/ 'white' (rather than /ˈbʲawɨ/). It's an orthography-influenced distinction that shouldn't be transcribed in IPA. Sol505000 (talk) 13:25, 25 April 2022 (UTC)

Consider the following pairs in their prescriptive pronunciation (not all speakers maintain the distinction, perhaps only a small minority, in part because the spelling reform of 1936 means there is no way to show it in spelling before any vowel other than i):

• (Widzę) lamie. / (Siedzę na) lamie.
• (Rozpoznaję) ksenofobię. / (Używam sobie na) ksenofobie.
• (Czytam) Utopię. / (Ja go chyba zaraz) utopię.
• (Opisuję) Batawię. / (Opowiadam o pewnym) Batawie.
• (Studiuję) filozofię. / (Rozmyślam o pewnym) filozofie.

In all these examples the first word ends (or used to end) in what a Russian would spell -ье, the second one -- -е (the vowel part of the example is of course problematic because of vowel reduction, the ь part is important).

(It's even more pronounced with -ii and -i (e.g. utopii -- -ьи́/ utopi -- -и́), but in this case the hypercorrection works the other way round (you can end up with Chylonii for Chyloni, konopii for konopi etc.).

How would you transcribe it? If "filozofie" is supposed to be /filɔˈzɔfjɛ/, then "filozofię" has to be something like /filɔˈzɔfi̯jɛ/, with non-syllabic i. It can't be /filɔˈzɔfijɛ/, that would be spelt "filozofije" (an archaism nowadays). Now, IPA is supposed to be a phonemic rather than phonetic alphabet -- good luck trying to convince anyone (even the few people who cling to the above distinction) that 1) there's an /i̯/ phoneme in Polish, 2) no, it's not the same thing as /j/. It's easier to transcribe the former as /fʲilɔˈzɔfʲɛ/, and the latter as /fʲilɔˈzɔfʲjɛ/. 89.64.69.33 (talk) 01:21, 23 September 2022 (UTC)

Sawicka in Fonologia w: Gramatyka współczesnego języka polskiego, Fonetyka i Fonologia. p.145 tackles this phenomenon as follows:

"Z. Stieber zachowując miękkie fonemy wargowe, powoływał sie na istnienie funkcjonalnej różnicy między sekwencjami typu */PʲV/ i /PjV/ (Stieber 1966, s. 108). Taką różnicę widział on na przykład między */mʲV/ i */mjV/ w np. ziemia i armia (*/#ʑɛmʲɑ#/ ale */#ɑrmjɑ#). Jednakże realizacja ostatniej sylaby w obu przypadkach jest w kulturalnej polszczyźnie identyczna. Jest to przykład czysto psychologicznego rozróżnienia, które w danym wypadku wzmacniane jest różnicą ortograficzna w przypadkach zależnych, por. armii [ɑrmʲi] vs ziemi [ʑemʲi]. U pewnej grupy osób występuje świadomość odrębności między ciągiem /PjV/, a hipotetycznym */PʲV/. Zróżnicowanie idzie tu po linii wyraz obcy vs wyraz rodzimy. Otóż w wyrazach rodzimych graficzna grupa PiV jest realizowana wyłącznie jako [PʲjV], np. biorę [bʲjɵrɛɯ̯̃], robię [rɔbʲjeɯ̯̃], podczas gdy w wyrazach obcych, obok wymienionej, zdarza się realizacja [PʲijV], np. biologia [bʲjɵlɔɟja] i [bʲijɵlɔɟja]. Osoby, które wymawiają w ten sposób, mają w tych wyrazach inną reprezentację fonologiczną: 1. w wyrazach rodzimych /PjV/ o obligatoryjnej realizacji [PʲjV] (według interpretacji starszej w miękkimi labialnymi fonemami: /PʲV/), 2. w wyrazach obcych /PiV/ o realizacji [PʲijV] zob. w rozdz. 3.1 (według starszej interpretacji /PʲjV/)."

"In retaining soft labial phonemes, Z. Stieber pointed out the existence of a functional difference between sequences of the type */ PʲV / and / PjV / (Stieber 1966, p. 108). He saw such a difference between */ mʲV / and */ mjV / e.g. in ziemia and armia (*/#ʑɛmʲɑ#/ but */#ɑrmjɑ#). However, the realization of the last syllable in both cases is identical in the cultural Polish language. This is an example of a purely psychological distinction, which in a given case is reinforced by a spelling difference in dependent cases, cf. armii [ɑrmʲi] vs ziemi [ʑemʲi]. A certain group of people is aware of the difference between the sequence /PjV/ and the hypothetical */PʲV/. Differentiation goes along the lines of the foreign word versus the native word. In native words, the grapheme PiV is realized only as [PʲjV], e.g. biorę [bʲjɵrɛɯ̯̃], robię [rɔbʲjeɯ̯̃], while in foreign words in can be realzied as [PʲijV] as well, e.g. biology [bʲjɵlɔɟja] and [bʲijɵlɔɟja]. People who pronounce them this way have a different phonological representation in these words: 1. in native words /PjV/ with obligatory realization [PʲjV] (according to the older interpretation with soft labial phonemes: /PʲV/), 2. in foreign words /PiV/ realized as [PʲijV] see. in chapter 3.1 (according to the older interpretation /PʲjV/)."

While the pronunciation of [PʲijV] is definitely dated now and is something like [Pʲj̆V] vs[PʲjV] (could be transcribed as well [PʲjV] vs [PʲjˑV], or even [PʲV] vs [PʲjV]), the system of those who still feel the difference can be accommodated within Sawicka's phonemicization without a need of extra phonemes. She includes in her list of phonemes a 'weak juncture' phoneme (noted below with a plus juncture /+/ rather than with her original /#̯/) to explain difference between e.g. objętość /#ɔb+jɛntɔɕt͡ɕ#/ [ɔbjentɔɕt͡ɕ] or [ɔbʲjentɔɕt͡ɕ] vs obiecać /#ɔbjɛt͡sɑt͡ɕ#/ [ɔbʲjet͡sɑt͡ɕ]. The same juncture can be used to explain '(Czytam) Utopię.' vs '(Ja go chyba zaraz) utopię.' as /#utɔp+jɛ(w̃)#/ vs /#utɔpjɛ(w̃)#/. This juncture is reinforced by a syllable break so alternatively /ɔb.jɛntɔɕt͡ɕ/ vs /ɔ.bjɛt͡sɑt͡ɕ/, /utɔp.jɛ(w̃)/ vs /utɔ.pjɛ(w̃)/. A similar solution is applied in English to differentiate between a name /ə.neɪm/ vs. an aim /ən.eɪm/, night rate /naɪt.reɪt/ vs. nitrate /naɪ.treɪt/ without resorting to claims of additional phonemes. Qerez (talk) 11:38, 1 November 2022 (UTC)